public class ClusterExample extends Object
All instances MUST use a different properties file, because their instance Ids must be different, however all other properties should be the same.
If you want it to clear out existing jobs & triggers, pass a command-line argument called "clearJobs".
You should probably start with a "fresh" set of tables (assuming you may have some data lingering in it from other tests), since mixing data from a non-clustered setup with a clustered one can be bad.
Try killing one of the cluster instances while they are running, and see that the remaining instance(s) recover the in-progress jobs. Note that detection of the failure may take up to 15 or so seconds with the default settings.
Also try running it with/without the shutdown-hook plugin registered with the scheduler. (org.quartz.plugins.management.ShutdownHookPlugin).
Note: Never run clustering on separate machines, unless their clocks are synchronized using some form of time-sync service (such as an NTP daemon).
SimpleRecoveryJob
,
SimpleRecoveryStatefulJob
Constructor and Description |
---|
ClusterExample() |
Copyright © 2013. All Rights Reserved.